105 Machine learning and carotid artery CT radiomics identify significant differences between culprit and non-culprit lesions in patients with stroke and transient ischaemic attack
Elizabeth Le, Nicholas R. Evans, Jason M. Tarkin, Mohammed M. Chowdhury, Fulvio Zaccagna, Holly Pavey, C. Wall, Yuan Huang, Jonathan Weir‐McCall, Elizabeth A. Warburton, Leonardo Rundo, Carola‐Bibiane Schönlieb, Evis Sala, James H.f. Rudd2020 · DOI
Carotid atherosclerosis is a primary driver of ischemic stroke. While clinical decisions often rely on the degree of stenosis, many strokes occur in patients with mild narrowing due to "vulnerable" plaques that are prone to rupture. Texture analysis, a radiomic approach, quantifies image heterogeneity to predict tissue characteristics. This study investigated whether radiomics could differentiate symptomatic from asymptomatic patients and culprit from non-culprit plaques using machine learning (ML) classification.
Methods
The study analyzed CT angiography (CTA) images from symptomatic patients with carotid-related cerebrovascular accidents (CVAs) and asymptomatic (ASX) controls. Regions-of-interest (ROIs) were manually segmented across 14 consecutive 3mm slices. The PyRadiomics library was used for isotropic resampling and normalization before extracting features from six classes, including first-order statistics and second-order texture features.
Radiomics pipeline: (A) ROI segmentation, (B) Mask generation, (C) 3D visualization, and (D) Voxel-based feature extraction.
To classify plaques as ASX, culprit (CC), or non-culprit (NC), features were processed through a Scikit-learn pipeline involving variance thresholding, scaling, and dimensionality reduction. A 5-fold cross-validation scheme was employed with five ML classifiers: Decision Tree, Naive Bayes, Random Forest, Logistic Regression, and Neural Network.
Results
The dataset included 82 carotid arteries from 41 symptomatic patients (41 CC, 41 NC) and 50 arteries from 25 asymptomatic patients. Statistical analysis revealed significant differences between CC and NC arteries in both first- and second-order features (e.g., GLDM Large Dependence High Grey-Level Emphasis, p<0.0001). Notably, both CC and NC plaques exhibited higher heterogeneity than ASX plaques (GLDM Dependence Entropy, p<0.0001), suggesting a systemic high-risk vascular state in stroke patients.
Multi-class classifier accuracy and 95% confidence intervals. The Neural Network achieved the highest performance.
All ML classifiers outperformed random guessing (33.3% accuracy). The Neural Network achieved the highest accuracy of 69% (CI [61%, 77%]) and an AUC of 0.82 (CI [0.78, 0.86]). Logistic Regression followed with an AUC of 0.80, while the Decision Tree performed the poorest.
Discussion and Conclusion
Textural analysis combined with machine learning reveals highly significant differences between symptomatic and asymptomatic patients, as well as between culprit and non-culprit arteries within the same patient. This approach moves beyond simple stenosis measurement toward a "virtual biopsy" that captures the internal complexity of plaque structure.
While limited by a relatively small retrospective dataset and manual segmentation, this study demonstrates strong potential for identifying high-risk patients. Future work should focus on larger multi-center cohorts and automated segmentation to facilitate clinical integration for stroke prevention and aggressive medical therapy surveillance.
Summary updated
Abstract
<h3>Introduction</h3> Carotid atherosclerosis is the main cause of ischaemic stroke. Texture analysis is a radiomic approach used to quantify image heterogeneity which can predict tumour aggression in oncology. We investigated whether this method could be applied to carotid artery disease to differentiate symptomatic from asymptomatic patients and culprit from non-culprit plaques, and then whether machine learning (ML) could correctly classify plaques based on these features. <h3>Methods</h3> CT angiography (CTA) images from symptomatic patients with carotid artery-related cerebrovascular accidents (CVAs) and from asymptomatic (ASX) patients were studied. Regions-of-interest (ROIs) were drawn on 14 consecutive carotid artery CTA slices with 3mm slice thickness. PyRadiomics was used for isotropic image (1x1x1) resampling and normalisation prior to texture feature extraction from 6 different classes (Table 1). Asymptomatic carotids were compared to culprit carotids (CC), and non-culprit (NC) carotids using the Mann Whitney U test or Wilcoxon signed-rank tests as appropriate, with a p-value <0.0005 deemed statistically significant after Bonferroni correction. Non-normally distributed variables are reported as median (interquartile range). To assess the discriminatory ability of radiomic features in multi-class classification (ASX, CC or NC), texture features were fed into a Python scikit-learn pipeline for feature selection with variance thresholding, feature scaling and dimensionality reduction incorporated within a 5-fold cross validation scheme with 5 different ML classifiers to reduce the risk of data leakage and overfitting. Mean cross validation (CV) accuracy, area under the receiver operating curve (AUC) and 95% confidence intervals (CI) are reported. <h3>Results</h3> The dataset comprised 82 carotid arteries from 41 symptomatic patients (41 culprit; 41 non-culprit) and 50 carotid arteries from 25 asymptomatic patients. CC and NC carotids showed significant differences in both first- and second-order features (IH Median: CC 618 (61); NC 646 (97), p<0.005) and (GLDM Large Dependence High Grey-Level Emphasis: CC 3147 (1837), NC 4811 (2181), p<0.0001), respectively. Both CC and NC carotids had higher heterogeneity than asymptomatic carotids (GLDM Dependence Entropy: CC 6.59 (0.43), NC 6.57 (0.52), ASX 6.24 (0.26), p<0.0001). All ML classifiers performed better than a randomly guessing classifier (mean accuracy 33.3%) in this multi-class (n=3) classification task (Table 2; Figure 2), with the neural network achieving the highest accuracy of 69%, CI [61%, 77%] with AUC 0.82 CI [0.78, 0.86]. <h3>Conclusions</h3> Textural analysis combined with machine learning on carotid CT scans reveals highly significant differences between symptomatic and asymptomatic patients, and between culprit and non-culprit carotid arteries within symptomatic patients. This approach could help identify patients at high-risk of stroke for aggressive medical therapy and surveillance. <h3>Conflict of Interest</h3> None
BibTeX
@misc{le2020machine,
author = {Elizabeth Le and Nicholas R. Evans and Jason M. Tarkin and Mohammed M. Chowdhury and Fulvio Zaccagna and Holly Pavey and C. Wall and Yuan Huang and Jonathan Weir‐McCall and Elizabeth A. Warburton and Leonardo Rundo and Carola‐Bibiane Schönlieb and Evis Sala and James H.F. Rudd},
title = {{105 Machine learning and carotid artery CT radiomics identify significant differences between culprit and non-culprit lesions in patients with stroke and transient ischaemic attack}},
year = {2020},
doi = {10.1136/heartjnl-2020-bcs.105},
}